Keeping a body going is actually much simpler than keeping a car going. Our bodies are actually quite well designed to replace their own parts, the parts are just much smaller.
To some degree certainly. Your body does regenerate, and has mechanisms to repair itself. These abilities degrade over time, but there's no reason you couldn't find some way to aid that process. The body can't do some things though. We can't grow a new organ. You blow out you hip, you don't grow a new one, etc.
Build machines small enough to kill individual bacterium and you've solved most of the worlds health problems (since those same machines could handle just about anything else as well). Cancer's not an issue if you have mini-machines picking off the cells as they develop.
You still have to have a way to identify your target. Not only find it and destroy it, but not follow false signals. Improvements to a delivery vector are important, and we really can't underscore that enough, but you still have to pick your targets somehow. Biomarker discovery is notoriously difficult process. It may be relatively simple to target a bacteria, you still have to make sure you're only targeting that particular strain, which is a bit more complicated. I mean our understanding of the human microbiome is still in it's infancy. Inadvertently killing off beneficial bacteria in your body could damage your immune system, cause you to overeat, lead to vitamin deficiency, and any number of other things.
In addition something like a cancer isn't just identifying the cancerous cells. You'll have a body with cells having all degrees of degradation from a lifetime of exposure to the environment. Catching them when they're sufficiently damaged to cause cancer, but not killing a cell that hasn't crossed that threshold would be difficult. There's much more of a grey area there than just "cancer" and "not cancer." In addition, not all cancers are necessarily dangerous anyway. Do we turn something loose in your body if we don't have to? How do we know when we've reached that point?
Nanotech certainly has a lot of potential as a delivery vector, but again, having the understanding of the human body that's necessary to choose the target is going to be the more difficult step. Edited, Jun 20th 2012 11:25am by someproteinguy